def test_fit_fill_value_mode_all_nulls(self):
"""If fill_value is mode and all the values are null, _fill_value is 0."""
# Setup
data = pd.Series([None, None, None])
# Run
transformer = NullTransformer(fill_value='mode')
transformer.fit(data)
# Asserts
assert transformer._fill_value == 0
def test_fit_fill_value(self):
"""If fill_value is something else, _fill_value is fill_value."""
# Setup
data = pd.Series([1, 2, 3, 3, 4])
# Run
transformer = NullTransformer(fill_value=43)
transformer.fit(data)
# Asserts
assert transformer._fill_value == 43
def test_fit_fill_value_mean_nulls(self):
"""If fill_value is mean, _fill_value is the mean of the input data."""
# Setup
data = pd.Series([1, None, 3])
# Run
transformer = NullTransformer(fill_value='mean')
transformer.fit(data)
# Asserts
assert transformer._fill_value == 2
def test_fit_fill_value_mode_nulls(self):
"""If fill_value is mode, _fill_value is the mode of the input data."""
# Setup
data = pd.Series([1, None, 3, 3, 4])
# Run
transformer = NullTransformer(fill_value='mode')
transformer.fit(data)
# Asserts
assert transformer._fill_value == 3
def test_fit_null_column_true(self):
"""Test fit with null_column equal to True"""
# Setup
data = pd.Series([1.5, None, 2.5])
# Run
fill_value = None
null_column = True
transformer = NullTransformer(fill_value, null_column=null_column)
transformer.fit(data)
# Asserts
self.assertTrue(transformer._null_column,
"_null_column must be True when null_column is True")
def test_fit_null_column_none_false(self):
"""Test fit with null_columns equal to None, no None values in data"""
# Setup
data = pd.Series([1.5, 3.0, 2.5])
# Run
fill_value = None
null_column = None
transformer = NullTransformer(fill_value, null_column=null_column)
transformer.fit(data)
# Asserts
self.assertFalse(
transformer._null_column,
"_null_column must be true when there are None values as input")
def test___init__(self):
"""Test default instance"""
# Run
transformer = NullTransformer(None)
# Asserts
self.assertIsNone(transformer.null_column,
"null_column is None by default")
self.assertFalse(transformer.copy, "copy is False by default")
def test_fit_null_column_none(self):
"""If null_column is None, _null_column is true only if there are Nulls."""
# Setup
nonulls = pd.Series([1, 2, 3])
nulls = pd.Series([1, None, 3])
# Run
nonulls_transformer = NullTransformer(fill_value=0, null_column=None)
nonulls_transformer.fit(nonulls)
nulls_transformer = NullTransformer(fill_value=0, null_column=None)
nulls_transformer.fit(nulls)
# Asserts
assert not nonulls_transformer._null_column
assert nulls_transformer._null_column
def test_fit_null_column_false(self):
"""If null_column is false, _null_column is always false."""
# Setup
nonulls = pd.Series([1, 2, 3])
nulls = pd.Series([1, None, 3])
# Run
nonulls_transformer = NullTransformer(fill_value=0, null_column=False)
nonulls_transformer.fit(nonulls)
nulls_transformer = NullTransformer(fill_value=0, null_column=False)
nulls_transformer.fit(nulls)
# Asserts
assert not nonulls_transformer._null_column
assert not nulls_transformer._null_column
def test_reverse_transform_no_nulls(self):
"""Test reverse_transform with no nulls"""
# Setup
data = pd.Series([1.5, None, 2.5])
# Run
transformer = Mock()
transformer.nulls = False
result = NullTransformer.reverse_transform(transformer, data)
# Asserts
expect = pd.Series([1.5, None, 2.5])
np.testing.assert_array_equal(result, expect,
"Unextected reverse data")
def test_transform_without_nulls(self):
"""Test transform data with nulls equal to False"""
# Setup
data = pd.Series([1.5, None, 2.5])
# Run
transformer = Mock()
transformer.nulls = False
result = NullTransformer.transform(transformer, data)
# Asserts
expect = np.array([1.5, np.nan, 2.5])
np.testing.assert_array_equal(result, expect,
"Unexpected transformed data")
def test_reverse_transform_nulls_and_null_column(self):
"""Test reverse_transform with nulls and null_column"""
# Setup
data = np.array([[1.5, 0], [0.6, 1], [2.5, 0]])
# Run
transformer = Mock()
transformer.nulls = True
transformer._null_column = True
result = NullTransformer.reverse_transform(transformer, data)
# Asserts
expect = pd.Series([1.5, np.nan, 2.5])
np.testing.assert_array_equal(result, expect,
"Unextected reverse data")
def test_transform_with_nulls_null_column(self):
"""Test transform data with nulls equal to True and null_column"""
# Setup
data = pd.Series([1.5, None, 2.5])
# Run
transformer = Mock()
transformer.nulls = True
transformer._fill_value = None
transformer._null_column = True
result = NullTransformer.transform(transformer, data)
# Asserts
expect = np.array([[1.5, 0], [np.nan, 1], [2.5, 0]])
np.testing.assert_array_equal(result, expect,
"Unexpected transformed data")
def test_transform_with_nulls_fillvalue_copy(self):
"""Test transform data with nulls equal to True and fill_value"""
# Setup
data = pd.Series([1.5, None, 2.5])
# Run
transformer = Mock()
transformer.nulls = True
transformer.copy = True
transformer.fill_value = 0
transformer._null_column = False
result = NullTransformer.transform(transformer, data)
# Asserts
expect = np.array([1.5, 0, 2.5])
np.testing.assert_array_equal(result, expect,
"Unexpected transformed data")
def test_transform_with_nulls_fillvalue_no_copy(self):
"""Test transform data with nulls equal to True and fill_value
NOTE: This test will execute a code that raise a warning.
We add the filterwarning to mute it since it's not a real warning in the test.
"""
# Setup
data = pd.Series([1.5, None, 2.5])
# Run
transformer = Mock()
transformer.nulls = True
transformer.copy = False
transformer._fill_value = 0
transformer._null_column = False
result = NullTransformer.transform(transformer, data)
# Asserts
expect = np.array([1.5, 0, 2.5])
np.testing.assert_array_equal(result, expect,
"Unexpected transformed data")
